[Pws] AlwaysCrashing
Hello, I have a problem with our swiki I cannot figure out. After years without problems our swiki regularly crashes every other day or even just hours after being started. I located the most recently changed swikis and turned individual logging on, but haven't found anything unusual afaik. (still have the logs if necessary) I ran it in gdb then and it seems that it tries to allocate _way_ too much memory and thus is killed by the OS. it crashes on both boxes i tried, under two different OSes, happily chewing 512MB RAM and 1.5GB swap. unfortunately the debugger is killed, too: Program received signal SIGKILL, Killed. 0x080638c2 in allocateheaderSizeh1h2h3doFillwith () (gdb) [1]+ Hangup: 1 squeak -headless Killed: 9 a lot of these kernel messages appear: swap_pager_getswapspace(10): failed and finally, pid 8214 (gdb), uid xxx, was killed: out of swap space Any ideas how to dig further to have it running nicely like earlier on ? tia, t ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws
Re: [Pws] AlwaysCrashing
Hi, I would increase the swap space to alot more and see if you can get another error message, but maybe some one else has a better solution to the issue. Good Luck! Date: Sun, 27 Jan 2008 14:38:31 +0100 From: [EMAIL PROTECTED] To: pws@cc.gatech.edu Subject: [Pws] AlwaysCrashing Hello, I have a problem with our swiki I cannot figure out. After years without problems our swiki regularly crashes every other day or even just hours after being started. I located the most recently changed swikis and turned individual logging on, but haven't found anything unusual afaik. (still have the logs if necessary) I ran it in gdb then and it seems that it tries to allocate _way_ too much memory and thus is killed by the OS. it crashes on both boxes i tried, under two different OSes, happily chewing 512MB RAM and 1.5GB swap. unfortunately the debugger is killed, too: Program received signal SIGKILL, Killed. 0x080638c2 in allocateheaderSizeh1h2h3doFillwith () (gdb) [1]+ Hangup: 1 squeak -headless Killed: 9 a lot of these kernel messages appear: swap_pager_getswapspace(10): failed and finally, pid 8214 (gdb), uid xxx, was killed: out of swap space Any ideas how to dig further to have it running nicely like earlier on ? tia, t ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws _ Helping your favorite cause is as easy as instant messaging. You IM, we give. http://im.live.com/Messenger/IM/Home/?source=text_hotmail_join___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws
Re: [Pws] AlwaysCrashing
ooh, sounds like a tough one! you might try turning on debugging and logDebuggerStackToFile in the image to see if you get any (squeak) stack traces related to this. is there any stack backtrace (where) available when you use gdb? are you running WinterLong? what is your disk space like on the partition where the logs are written? (i was getting daily crashes a while back and it turned out my logs were filling up the partition) [scratches his head] hal On Jan 27, 2008, at 6:38 AM, sa9k063 wrote: Hello, I have a problem with our swiki I cannot figure out. After years without problems our swiki regularly crashes every other day or even just hours after being started. I located the most recently changed swikis and turned individual logging on, but haven't found anything unusual afaik. (still have the logs if necessary) I ran it in gdb then and it seems that it tries to allocate _way_ too much memory and thus is killed by the OS. it crashes on both boxes i tried, under two different OSes, happily chewing 512MB RAM and 1.5GB swap. unfortunately the debugger is killed, too: Program received signal SIGKILL, Killed. 0x080638c2 in allocateheaderSizeh1h2h3doFillwith () (gdb) [1]+ Hangup: 1 squeak -headless Killed: 9 a lot of these kernel messages appear: swap_pager_getswapspace(10): failed and finally, pid 8214 (gdb), uid xxx, was killed: out of swap space Any ideas how to dig further to have it running nicely like earlier on ? tia, t ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws
Re: [Pws] AlwaysCrashing
meant to say from the core file--i realize that gdb is crashing as well, so you can't see this when it happens. i'm also not sure which one will create the core file). is there any stack backtrace (where) available when you use gdb? ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws
Re: [Pws] AlwaysCrashing
We had something like that happen recently. We found that the log file was at 2Gb which turns out to be some kind of Squeak limit. We dumped it and started over, and all was well. Mark -Original Message- From: [EMAIL PROTECTED] on behalf of [EMAIL PROTECTED] Sent: Sun 1/27/2008 11:53 AM To: pws@cc.gatech.edu Subject: Re: [Pws] AlwaysCrashing Hi, I would increase the swap space to alot more and see if you can get another error message, but maybe some one else has a better solution to the issue. Good Luck! Date: Sun, 27 Jan 2008 14:38:31 +0100 From: [EMAIL PROTECTED] To: pws@cc.gatech.edu Subject: [Pws] AlwaysCrashing Hello, I have a problem with our swiki I cannot figure out. After years without problems our swiki regularly crashes every other day or even just hours after being started. I located the most recently changed swikis and turned individual logging on, but haven't found anything unusual afaik. (still have the logs if necessary) I ran it in gdb then and it seems that it tries to allocate _way_ too much memory and thus is killed by the OS. it crashes on both boxes i tried, under two different OSes, happily chewing 512MB RAM and 1.5GB swap. unfortunately the debugger is killed, too: Program received signal SIGKILL, Killed. 0x080638c2 in allocateheaderSizeh1h2h3doFillwith () (gdb) [1]+ Hangup: 1 squeak -headless Killed: 9 a lot of these kernel messages appear: swap_pager_getswapspace(10): failed and finally, pid 8214 (gdb), uid xxx, was killed: out of swap space Any ideas how to dig further to have it running nicely like earlier on ? tia, t ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws Helping your favorite cause is as easy as instant messaging. You IM, we give. Learn more. http://im.live.com/Messenger/IM/Home/?source=text_hotmail_join ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws
Re: [Pws] AlwaysCrashing
Guzdial, Mark wrote: We had something like that happen recently. We found that the log file was at 2Gb which turns out to be some kind of Squeak limit. We dumped it and started over, and all was well. Mark I just checked, the logfiles are all less than 10MB in size. would have been great to have it fixed so easily, though :) greets, tee ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws
Re: [Pws] AlwaysCrashing
Hal Eden wrote: ooh, sounds like a tough one! you might try turning on debugging and logDebuggerStackToFile in the image to see if you get any (squeak) stack traces related to this. cool, thanks, that sounds like i could get more debug info. while i'm at it, should i enable securityChecksEnabled and/or cpuWatcherEnabled ? after the swiki is started an running a while cpu load is at 10..20%. before the squeak process crashes, cpu utilization is right under the ceiling. is there any stack backtrace (where) available when you use gdb? the only line i got was 0x080638c2 in allocateheaderSizeh1h2h3doFillwith () are you running WinterLong? it currently is Squeak 3.7-7 #5989 (bsd ports) :: Comanche 7.0.2 :: Swiki 1.5 on FreeBSD (but has earlier been running on macOS9, OSX and linux as well). the symptoms started on the linux box, so i put it onto a bsd box where it ran for a week before the daily crashing happened again. i initially activated the image to start on a mac and copied it to the server (no X forwarding from home and no X on the server) and started it there. what is your disk space like on the partition where the logs are written? (i was getting daily crashes a while back and it turned out my logs were filling up the partition) there is still plenty of space, on its original box i leave it running on a partition with ~10 Gigs of free space, on the box now it is even more. (it has less ram, though) i just started it with -memory 128m -mmap 512m -headless to see if it makes a difference. [scratches his head] hal cheers, tee ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws
Re: [Pws] AlwaysCrashing
On Mon, Jan 28, 2008 at 12:05:07AM +0100, sa9k063 wrote: [EMAIL PROTECTED] wrote: Hi, I would increase the swap space to alot more and see if you can get another error message, but maybe some one else has a better solution to the issue. i can try that someday next week, the other machine has a second disk i could add swap with. If you are swapping to disk, you are already in trouble. Squeak will slow to a crawl if it has to start swapping. Is the Squeak image growing too large for some other reason perhaps? Dave ___ Pws mailing list Pws@cc.gatech.edu https://mailman.cc.gatech.edu/mailman/listinfo/pws